DNA microarrays have become an extremely important bioanalytical tool (e.g. for analyzing gene expression); protein microarray technology has, however, lagged behind. The fabrication of protein arrays is challenging because of difficulties associated with preserving the folded conformation of proteins in the immobilized state, and high amounts of non-specific binding to immobilized proteins. As a large fraction of drug targets are membrane bound proteins (e.g., G-protein coupled receptors, ion-channels, etc.), there is an impetus to develop tools for high-throughput screening against membrane bound proteins. Membrane proteins maintain their folded conformation when associated with lipids; therefore, to create arrays of such proteins it is important to first develop surfaces that support the binding of membranes. Bilayer-lipid membranes adsorbed onto solid supports, referred to as supported bilayer-lipid membranes, can mimic the structural and functional role of biological membranes. (See Sackmann, E. Science 1996, 271, 43-48; Bieri, C. et al., Nature Biotech, 1999, 17, 1105-1108; Groves, J. T. et al., Science 1997, 275, 651-653; Lang, H. et al., Langmuir 1994, 10, 197-210; Plant, A. L. et al., Langmuir 1999, 15, 5128-5135; and Raguse, B. et al., Langmuir 1998, 14, 648-659.) These hybrid surfaces were developed to overcome the fragility of black lipid membranes while preserving aspects of lateral fluidity observed in native biological membranes.
Surfaces that bind lipid membranes can be broadly classified into three categories:                (i) hydrophobic surfaces (e.g., self-assembled monolayers presenting terminal methyl groups) that support the adsorption of lipid monolayers; these are of limited utility as they cannot be used to incorporate membrane-spanning proteins (Plant, A. L., Langmuir 1999, 15, 5128-5135);        (ii) hydrophilic surfaces (e.g., glass surfaces) that bind bilayer-lipid membranes; these are also of limited utility as they can only be used to incorporate membrane-spanning proteins with extra-membrane domains that are less thick than the layer of adsorbed water (˜1 nm) (Groves, J. T. et al., Science 1997, 275, 651-653; and Groves, J. T. et al., Langmuir 1998, 14, 3347-3350); and        (iii) amphiphilic surfaces that contain hydrophobic and hydrophilic portions and bind bilayer-lipid membranes; these offer the potential for incorporating a wide variety of membrane-spanning proteins (Lang, H. et al., Langmuir 1994, 10, 197-210; Raguse, B. et al., Langmuir 1998, 14, 648-659; and Vanderah, D. J. et al., Materials Research Society Fall Meeting Abstracts, Boston, 1999).        
Methods to create arrays of membranes would enable high-throughput screening of multiple targets against multiple drug-candidates. Arrays of membranes may be obtained by fabricating grids of titanium oxide on a glass substrate as titanium oxide resists the adsorption of lipids (Boxer, S. G. et al. Science 1997, 275, 651-653; and Boxer, S. G. et al. Langmuir 1998, 14, 3347-3350). Micropipeting techniques have been used to spatially address each corralled lipid-binding region (Cremer, P. S. et al., J. Am. Chem. Soc. 1999, 121, 8130-8131). However, these methods are cumbersome and require the fabrication of patterned surfaces. To make membrane arrays by printing membranes on unpatterned surfaces, it would be necessary to confine the membrane to the printed areas without lateral diffusion of the membrane molecules to the unprinted areas. Boxer et al. have demonstrated that it is possible to pattern lipids on glass surfaces by microcontact printing using poly-dimethylsiloxane (PDMS) stamps “inked” with phosphatidylcholine (PC). They attributed the lateral confinement of the lipids to the stamped regions, to the self-limiting expansion of PC membranes to ˜106% of the original printed areas (Hovis, J. et al., Langmuir 2000, 16, 894-897). The methods used by Boxer et al., however, have certain limitations. First, Boxer and co-workers carried out the stamping of lipids on surfaces immersed under water (Hovis 2000). Second, lipids adsorbed on the bare-glass substrates used by Boxer and co-workers spontaneously desorbed when drawn through an air-water interface (Cremer 1999). Cremer et al., propose in WO01/20330 the use of spatially addressed lipid bilayer arrays that remain submerged underwater to preserve the planar support structure. Such systems may not be practical for robust, high throughput, microarray based assays.